The invention discloses a big data analysis method based on intermittent faults, and belongs to the electric power equipment technology field. According to the prior art, determining of an abnormity emergency degree is mainly dependent on human experience and subjective estimation, and order sending processing is carried out, and the above mentioned method is random, and is not scientific, and therefore worker dispatching wastes can be caused, and normal worker dispatching is delayed, and a scientific, reasonable, and accurate big data analysis scheme is required. The influence factors of the acquisition abnormity emergency degree are determined, and comprise the intermittent faults. Data comprises on-site acquisition data and work order flowing data, and then fluctuation analysis of electricity utilization quantity of a lot of users is carried out. The influence factors are analyzed, and models are established, and complicated problems are systematized, stratified, and simplified, and therefore a feasible technical scheme is provided. The influence factors are identified and discriminated, and the influence factors having the influences on the emergency degree are brought into an emergency degree calculation model, and the influence factors having no influences on the emergency degree are not brought into the emergency degree calculation model, and therefore the worker dispatching wastes are prevented.